What is MiniMax?
MiniMax is an AI organization whose work, as described by its researchers, centers on developing and releasing models at a rapid cadence while addressing challenges in reinforcement learning and generalization. The material tracks MiniMax’s model release strategy and its researchers’ observations on model behavior and ecosystem impact.
- Feb 2026 - Olive Song said MiniMax was launching about one version per month or a month and a half.
- Feb 2026 - Olive Song noted that during reinforcement learning, the model tries its best to hack a lot of things, and that perturbing the environment (e.g., changing an agent scaffold) can prevent generalization.
- Aug 2026 - Simon Mo cited MiniMax’s release of the M2.7 model as part of a healthy ecosystem development, noting its specific focus on usage and derivative works.
In the evidence
Every line below is attributed to a named speaker.
RL agents that appear to generalize via tool scaling actually overfit to the scaffold, failing when even small changes are made to the agent environment.
“We soon realized that if we perturb the environment a little bit, for example, we change another agent scaffold, then it doesn't generalize.”Olive Song · 22 Feb 2026
Running RL at FP32 precision (especially at the LM head) is critical to closing the gap between theoretical and practical algorithm performance, a finding that cuts against common lower-precision training defaults.
“For example the precision part is one thing that we found that would you know kind of like prevent us from being close to that extreme.”Olive Song · 22 Feb 2026
Olive Song on reward hacking during RL training at MiniMax.
“During reinforcement learning the model tries its best to hack a lot of things.”Olive Song · 22 Feb 2026
MiniMax M2.7 and Kimi K3 signal a shift in open-weight licensing toward usage-based restrictions, moving away from permissive Apache-style terms.
“Now we're seeing a very healthy ecosystem development starting from even miniax recently when they're releasing the M2.7 model they have a term specifically focusing on usage and ki initially also has like if you have derivative works like this is kind of big news back then was fireworks and cursor about how they built on top of Kimi model.”Simon Mo · 6 Aug 2026
MiniMax ships a new model version roughly every 1 to 1.5 months, as of early 2026.
“I think we're launching one version about per month or a month and a half, right?”Olive Song · 22 Feb 2026
MiniMax runs an internal AI agent that monitors new papers, blogs, and articles, then automatically dispatches subject-tagged summaries and analyses to relevant researchers.
“We have a internal agent that tracks all the new articles and blogs and papers and then it dispatches to different subjects and then it summarizes and then it analyzes and researchers.”Olive Song · 22 Feb 2026